In certain computing architectures, such as “cloud computing” or other instances where architecture is provided as a service, security between different clients may be an important concern. For example, clients may pay for a certain number of cores and a certain volume of memory. Cloud computing means that the client need not have the same cores and the same memory every time. They can be dynamically allocated as needed. However, when memory is allocated from one client to another, it may need to be completely cleared to ensure that there is no data leakage from one client to the other. This can take the form of writing all 1's, all 0's, or a random or pseudorandom sequence to memory to ensure that any old values are no longer usable or recoverable.